-
@mohamed82008 and I had a council meeting before the next release, and we reconsidered the question of nondifferentiated positional arguments, introduced in #89 by @thorek1 (thanks for the contributio…
-
Using Enzyme `v0.11.1` and `main` with Julia 1.9.1 on AMD, Intel and M2 CPU gives erroneous results when using `autodiff_deferred` in Reverse mode in a kernel that has an outer-loop decorated with `Th…
-
MWE:
```julia
using Enzyme, SpecialFunctions;
first(autodiff(Reverse, erf, Active(0.2)));
```
-
Forgive me if this is a known issue, feel free to close if so. I see some stuff with complex numbers has been merged recently, so I tried this on main (8784d1f) with the same issue.
I'm trying to t…
-
### Description
Unless I am mistaken, differentiating the BCOO constructor appears to produce dense matrices. Here's an MWE:
```python
import jax
import jax.numpy as jnp
from jax.experimenta…
-
This works on Julia 1.9.2 and Enzyme main (b35703b):
```julia
using Enzyme
f(x) = sum(Set([1.0, x, 2x, x]))
autodiff(Reverse, f, Active, Active(2.0))[1][1] # 3.0
autodiff(Forward, f, Duplicated(2…
-
After @vchuravy's amazing PASC talk, I was convinced to finally take the plunge and convert my package, ElectrochemicalKinetics.jl, to Enzyme – it does a lot of scalar AD so I anticipate a nice perfor…
-
Hi,
I got a bug when running the following code
```julia
using Enzyme
using LinearAlgebra
using Optimization, OptimizationOptimJL
n = 4
function mylsovle(b::Vector{Float64}, A::Matrix{Floa…
-
I'm hitting [Enzyme.jl #644](https://github.com/EnzymeAD/Enzyme.jl/issues/644) via `optprob = OptimizationFunction(foo, Optimization.AutoEnzyme(), cons = cons)`. The Enzyme fix is not able to land on…
-
I am on Julia 1.9.0 and Enzyme 9487eb8349fd7d907403533b53c23822313897bb. This code errors, forward mode works.
```julia
using Enzyme
function f(x)
y = [1.0, 2.0]
empty!(y)
return x
…